In [1]:
import matplotlib.image as mpimg
import matplotlib.pyplot as plt
import numpy as np
import cv2
import glob
import time
import pickle
from sklearn import svm
from sklearn.svm import LinearSVC
from sklearn.preprocessing import StandardScaler
from skimage.feature import hog
from scipy.ndimage.measurements import label
from sklearn.cross_validation import train_test_split
%matplotlib inline
load data¶
In [2]:
images = glob.glob('*vehicles/*/*.png')
cars = []
notcars = []
for image in images:
if 'non' in image:
notcars.append(image)
else:
cars.append(image)
print(len(cars), len(notcars))
In [3]:
f, axs = plt.subplots(6, 6, figsize=(20,20))
axs = axs.ravel()
for i in range(18):
img = cv2.imread(cars[np.random.randint(0, len(cars))])
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
axs[i].axis('off')
axs[i].set_title('car', fontsize=10)
axs[i].imshow(img)
print('\n')
for i in range(18,36):
img = cv2.imread(notcars[np.random.randint(0, len(notcars))])
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
axs[i].axis('off')
axs[i].set_title('not car', fontsize=10)
axs[i].imshow(img)
plt.savefig('./output_images/dataset_visualize'+'.jpg')
In [4]:
# Define a function to return some characteristics of the dataset
def data_look(car_list, notcar_list):
data_dict = {}
# Define a key in data_dict "n_cars" and store the number of car images
data_dict["n_cars"] = len(car_list)
# Define a key "n_notcars" and store the number of notcar images
data_dict["n_notcars"] = len(notcar_list)
# Read in a test image, either car or notcar
test_img = mpimg.imread(car_list[0])
# Define a key "image_shape" and store the test image shape 3-tuple
data_dict["image_shape"] = test_img.shape
# Define a key "data_type" and store the data type of the test image.
data_dict["data_type"] = test_img.dtype
# Return data_dict
return data_dict
data_info = data_look(cars, notcars)
print('Your function returned a count of',
data_info["n_cars"], ' cars and',
data_info["n_notcars"], ' non-cars')
print('of size: ',data_info["image_shape"], ' and data type:',
data_info["data_type"])
# Just for fun choose random car / not-car indices and plot example images
car_ind = np.random.randint(0, len(cars))
notcar_ind = np.random.randint(0, len(notcars))
# Read in car / not-car images
car_image = mpimg.imread(cars[car_ind])
notcar_image = mpimg.imread(notcars[notcar_ind])
# # Plot the examples
# fig = plt.figure()
# plt.subplot(121)
# plt.imshow(car_image)
# plt.title('Example Car Image')
# plt.subplot(122)
# plt.imshow(notcar_image)
# plt.title('Example Not-car Image')
explore different color space¶
In [5]:
# Define a function to return HOG features and visualization
def get_hog_features(img, orient, pix_per_cell, cell_per_block, vis=False, feature_vec=True):
if vis == True:
# Use skimage.hog() to get both features and a visualization
#features = [] # Remove this line
features, hog_image = hog(img, orientations=orient, pixels_per_cell=(pix_per_cell, pix_per_cell), cells_per_block=(cell_per_block, cell_per_block), visualise=True, feature_vector=False)
return features, hog_image
else:
# Use skimage.hog() to get features only
features= hog(img, orientations=orient, pixels_per_cell=(pix_per_cell, pix_per_cell), cells_per_block=(cell_per_block, cell_per_block), visualise=False, feature_vector=False)
return features
# Generate a random index to look at a car image
ind = np.random.randint(0, len(cars))
# Read in the image
image = mpimg.imread(cars[ind])
gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
# Define HOG parameters
orient = 9
pix_per_cell = 8
cell_per_block = 2
# Call our function with vis=True to see an image output
features, hog_image = get_hog_features(gray, orient,
pix_per_cell, cell_per_block,
vis=True, feature_vec=False)
# Generate a random index to look at a car image
ind = np.random.randint(0, len(cars))
# Read in the image
not_image = mpimg.imread(notcars[ind])
gray = cv2.cvtColor(not_image, cv2.COLOR_RGB2GRAY)
features, not_hog_image = get_hog_features(gray, orient,
pix_per_cell, cell_per_block,
vis=True, feature_vec=False)
# Plot the examples
fig,((ax1, ax2),(ax3, ax4)) = plt.subplots(2, 2, figsize=(10,10))
ax1.imshow(image, cmap='gray')
ax1.set_title('Example Car Image')
ax2.imshow(hog_image, cmap='gray')
ax2.set_title('HOG Visualization')
ax3.imshow(not_image, cmap='gray')
ax3.set_title('Example Not Car Image')
ax4.imshow(not_hog_image, cmap='gray')
ax4.set_title('HOG Visualization')
plt.savefig('./output_images/dataset_hog_example'+'.jpg')
In [6]:
# Save the parameters for later use
dist_pickle = {}
# Define a key in data_dict "n_cars" and store the number of car images
dist_pickle["n_cars"] = len(cars)
# Define a key "n_notcars" and store the number of notcar images
dist_pickle["n_notcars"] = len(notcars)
# Read in a test image, either car or notcar
test_img = mpimg.imread(cars[0])
# Define a key "image_shape" and store the test image shape 3-tuple
dist_pickle["image_shape"] = test_img.shape
# Define a key "data_type" and store the data type of the test image.
dist_pickle["data_type"] = test_img.dtype
pickle.dump(dist_pickle, open("./data_pickle/data_dist_pickle.p", "wb"))
define functions of combine features¶
In [7]:
# Define a function to compute binned color features
def bin_spatial(img, size=(32, 32)):
# Use cv2.resize().ravel() to create the feature vector
features = cv2.resize(img, size).ravel()
# Return the feature vector
return features
# Define a function to compute color histogram features
# NEED TO CHANGE bins_range if reading .png files with mpimg!
def color_hist(img, nbins=32, bins_range=(0, 256)):
# Compute the histogram of the color channels separately
channel1_hist = np.histogram(img[:,:,0], bins=nbins, range=bins_range)
channel2_hist = np.histogram(img[:,:,1], bins=nbins, range=bins_range)
channel3_hist = np.histogram(img[:,:,2], bins=nbins, range=bins_range)
# Concatenate the histograms into a single feature vector
hist_features = np.concatenate((channel1_hist[0], channel2_hist[0], channel3_hist[0]))
# Return the individual histograms, bin_centers and feature vector
return hist_features
# Define a function to extract features from a list of images
# Have this function call bin_spatial() and color_hist()
def extract_features(imgs, color_space='RGB', spatial_size=(32, 32),
hist_bins=32, orient=9,
pix_per_cell=8, cell_per_block=2, hog_channel=0,
spatial_feat=True, hist_feat=True, hog_feat=True):
# Create a list to append feature vectors to
features = []
# Iterate through the list of images
for file in imgs:
file_features = []
# Read in each one by one
image = mpimg.imread(file)
# apply color conversion if other than 'RGB'
if color_space != 'RGB':
if color_space == 'HSV':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2HSV)
elif color_space == 'LUV':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2LUV)
elif color_space == 'HLS':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2HLS)
elif color_space == 'YUV':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2YUV)
elif color_space == 'YCrCb':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2YCrCb)
else: feature_image = np.copy(image)
if spatial_feat == True:
spatial_features = bin_spatial(feature_image, size=spatial_size)
file_features.append(spatial_features)
if hist_feat == True:
# Apply color_hist()
hist_features = color_hist(feature_image, nbins=hist_bins)
file_features.append(hist_features)
if hog_feat == True:
# Call get_hog_features() with vis=False, feature_vec=True
if hog_channel == 'ALL':
hog_features = []
for channel in range(feature_image.shape[2]):
hog_features.append(get_hog_features(feature_image[:,:,channel],
orient, pix_per_cell, cell_per_block,
vis=False, feature_vec=True))
hog_features = np.ravel(hog_features)
else:
hog_features = get_hog_features(feature_image[:,:,hog_channel], orient,
pix_per_cell, cell_per_block, vis=False, feature_vec=True)
# Append the new feature vector to the features list
file_features.append(hog_features)
features.append(np.concatenate(file_features))
# Return list of feature vectors
return features
print('Functions for extracting combined features done.')
extract features¶
In [8]:
color_space = 'YCrCb' # Can be RGB, HSV, LUV, HLS, YUV, YCrCb
orient = 9 # HOG orientations
pix_per_cell = 8 # HOG pixels per cell
cell_per_block = 2 # HOG cells per block
hog_channel = "ALL" # Can be 0, 1, 2, or "ALL"
spatial_size = (16, 16) # Spatial binning dimensions
hist_bins = 32 # Number of histogram bins
spatial_feat = True # Spatial features on or off
hist_feat = True # Histogram features on or off
hog_feat = True # HOG features on or off
visualize = True # Visualize hog image on or off
y_start_stop = [400, 656] # Min and max in y to search in slide_window()
scale = 1.5 # A parameter for the function finding cars
t = time.time()
car_features = extract_features(cars, color_space=color_space,
spatial_size=spatial_size, hist_bins=hist_bins,
orient=orient, pix_per_cell=pix_per_cell,
cell_per_block=cell_per_block,
hog_channel=hog_channel, spatial_feat=spatial_feat,
hist_feat=hist_feat, hog_feat=hog_feat)
notcar_features = extract_features(notcars, color_space=color_space,
spatial_size=spatial_size, hist_bins=hist_bins,
orient=orient, pix_per_cell=pix_per_cell,
cell_per_block=cell_per_block,
hog_channel=hog_channel, spatial_feat=spatial_feat,
hist_feat=hist_feat, hog_feat=hog_feat)
t2 = time.time()
print(round(t2-t, 2), 'Seconds to extract Hog|bin spatial|histogram features...')
normalize features and split dataset¶
In [9]:
# Create an array stack of feature vectors
X = np.vstack((car_features, notcar_features)).astype(np.float64)
# Fit a per-column scaler
# Compute the mean and std to be used for later scaling.
X_scaler = StandardScaler().fit(X)
# Apply the scaler to X
# Perform standardization by centering and scaling
scaled_X = X_scaler.transform(X)
# Define the labels vector
y = np.hstack((np.ones(len(car_features)), np.zeros(len(notcar_features))))
# Split up data into randomized training and test sets
rand_state = np.random.randint(0, 100)
X_train, X_test, y_train, y_test = train_test_split(
scaled_X, y, test_size=0.2, random_state=rand_state)
In [10]:
print('Parameters: \n',
'\n feature vector length X_train: ', len(X_train[0]),
'\n Scaled_X_train: ', len(scaled_X[0]),
'\n Color Space: ',color_space,
'\n orient: ',orient,
'\n pixels per cell: ', pix_per_cell,
'\n cells per block: ', cell_per_block,
'\n hog_channel: ', hog_channel,
'\n spatial_size: ', spatial_size,
'\n hist_bins: ', hist_bins,
'\n ystart: ', y_start_stop[0],
'\n ystop: ', y_start_stop[1])
Classify¶
In [11]:
### Classify
from sklearn import svm, grid_search
from sklearn.grid_search import GridSearchCV
# Use a linear SVC
# Automatically tuning parameters
# # Check the training time for the SVC
# t=time.time()
# parameters = {'kernel':('linear', 'rbf'), 'C':[1, 10]}
# svr = svm.SVC()
# svc = grid_search.GridSearchCV(svr, parameters)
# svc.fit(X_train, y_train)
# t2 = time.time()
# Check the training time for the SVC
t=time.time()
svc = LinearSVC()
svc.fit(X_train, y_train)
t2 = time.time()
print(round(t2-t, 2), 'Seconds to train SVC using GridSearchCV')
# Check the score of the SVC
print('Test Accuracy of SVC = ', round(svc.score(X_test, y_test), 4))
# Check the prediction time for a single sample
t=time.time()
n_predict = 10
print('My SVC predicts: ', svc.predict(X_test[0:n_predict]))
print('For these',n_predict, 'labels: ', y_test[0:n_predict])
t2 = time.time()
print(round(t2-t, 5), 'Seconds to predict', n_predict,'labels with SVC')
sliding windows¶
In [12]:
# Define a function that takes an image,
# start and stop positions in both x and y,
# window size (x and y dimensions),
# and overlap fraction (for both x and y)
def slide_window(img, x_start_stop=[None, None], y_start_stop=[None, None],
xy_window=(64, 64), xy_overlap=(0.5, 0.5)):
# If x and/or y start/stop positions not defined, set to image size
if x_start_stop[0] == None:
x_start_stop[0] = 0
if x_start_stop[1] == None:
x_start_stop[1] = img.shape[1]
if y_start_stop[0] == None:
y_start_stop[0] = 0
if y_start_stop[1] == None:
y_start_stop[1] = img.shape[0]
# Compute the span of the region to be searched
xspan = x_start_stop[1] - x_start_stop[0]
yspan = y_start_stop[1] - y_start_stop[0]
# Compute the number of pixels per step in x/y
nx_pix_per_step = np.int(xy_window[0]*(1 - xy_overlap[0]))
ny_pix_per_step = np.int(xy_window[1]*(1 - xy_overlap[1]))
# Compute the number of windows in x/y
nx_buffer = np.int(xy_window[0]*(xy_overlap[0]))
ny_buffer = np.int(xy_window[1]*(xy_overlap[1]))
nx_windows = np.int((xspan-nx_buffer)/nx_pix_per_step)
ny_windows = np.int((yspan-ny_buffer)/ny_pix_per_step)
# Initialize a list to append window positions to
window_list = []
# Loop through finding x and y window positions
# Note: you could vectorize this step, but in practice
# you'll be considering windows one by one with your
# classifier, so looping makes sense
for ys in range(ny_windows):
for xs in range(nx_windows):
# Calculate window position
startx = xs*nx_pix_per_step + x_start_stop[0]
endx = startx + xy_window[0]
starty = ys*ny_pix_per_step + y_start_stop[0]
endy = starty + xy_window[1]
# Append window position to list
window_list.append(((startx, starty), (endx, endy)))
# Return the list of windows
return window_list
# Define a function to draw bounding boxes
def draw_boxes(img, bboxes, color=(0, 0, 255), thick=6):
# Make a copy of the image
imcopy = np.copy(img)
# Iterate through the bounding boxes
for bbox in bboxes:
# Draw a rectangle given bbox coordinates
cv2.rectangle(imcopy, bbox[0], bbox[1], color, thick)
# Return the image copy with boxes drawn
return imcopy
In [13]:
def single_img_features(img, color_space='RGB', spatial_size=(32, 32),
hist_bins=32, orient=9,
pix_per_cell=8, cell_per_block=2, hog_channel=0,
spatial_feat=True, hist_feat=True, hog_feat=True):
#1) Define an empty list to receive features
img_features = []
#2) Apply color conversion if other than 'RGB'
if color_space != 'RGB':
if color_space == 'HSV':
feature_image = cv2.cvtColor(img, cv2.COLOR_RGB2HSV)
elif color_space == 'LUV':
feature_image = cv2.cvtColor(img, cv2.COLOR_RGB2LUV)
elif color_space == 'HLS':
feature_image = cv2.cvtColor(img, cv2.COLOR_RGB2HLS)
elif color_space == 'YUV':
feature_image = cv2.cvtColor(img, cv2.COLOR_RGB2YUV)
elif color_space == 'YCrCb':
feature_image = cv2.cvtColor(img, cv2.COLOR_RGB2YCrCb)
else: feature_image = np.copy(img)
#3) Compute spatial features if flag is set
if spatial_feat == True:
spatial_features = bin_spatial(feature_image, size=spatial_size)
#4) Append features to list
img_features.append(spatial_features)
#print(spatial_features.shape)
#5) Compute histogram features if flag is set
if hist_feat == True:
hist_features = color_hist(feature_image, nbins=hist_bins)
#6) Append features to list
img_features.append(hist_features)
#print(hist_features.shape)
#7) Compute HOG features if flag is set
if hog_feat == True:
if hog_channel == 'ALL':
hog_features = []
for channel in range(feature_image.shape[2]):
hog_features.append(get_hog_features(feature_image[:,:,channel],
orient, pix_per_cell, cell_per_block,
vis=False, feature_vec=True))
hog_features = np.ravel(hog_features)
else:
hog_features = get_hog_features(feature_image[:,:,hog_channel], orient,
pix_per_cell, cell_per_block, vis=False, feature_vec=True)
#8) Append features to list
img_features.append(hog_features)
#9) Return concatenated array of features
return np.concatenate(img_features)
# Define a function you will pass an image
# and the list of windows to be searched (output of slide_windows())
def search_windows(img, windows, clf, scaler, color_space='RGB',
spatial_size=(32, 32), hist_bins=32,
hist_range=(0, 256), orient=9,
pix_per_cell=8, cell_per_block=2,
hog_channel=0, spatial_feat=True,
hist_feat=True, hog_feat=True):
#1) Create an empty list to receive positive detection windows
on_windows = []
#2) Iterate over all windows in the list
for window in windows:
#3) Extract the test window from original image
test_img = cv2.resize(img[window[0][1]:window[1][1], window[0][0]:window[1][0]], (64, 64))
#4) Extract features for that window using single_img_features()
features = single_img_features(test_img, color_space=color_space,
spatial_size=spatial_size, hist_bins=hist_bins,
orient=orient, pix_per_cell=pix_per_cell,
cell_per_block=cell_per_block,
hog_channel=hog_channel, spatial_feat=spatial_feat,
hist_feat=hist_feat, hog_feat=hog_feat)
#5) Scale extracted features to be fed to classifier
test_features = scaler.transform(np.array(features).reshape(1, -1))
#6) Predict using your classifier
prediction = clf.predict(test_features)
#7) If positive (prediction == 1) then save the window
if prediction == 1:
on_windows.append(window)
#8) Return windows for positive detections
return on_windows
test sliding window¶
In [19]:
images = glob.glob('test_images/*.jpg')
window_list = []
y_start_stop = [400, 656]
for file_image in images:
image=cv2.imread(file_image)
image=cv2.cvtColor(image, cv2.COLOR_BGR2RGB) # Equal to mpimg.imread(file_image)
draw_image = np.copy(image)
#image = draw_image.astype(np.float32)/255
# Uncomment the following line if you extracted training
# data from .png images (scaled 0 to 1 by mpimg) and the
# image you are searching is a .jpg (scaled 0 to 255)
image = draw_image.astype(np.float32)/255
windows = slide_window(image, x_start_stop=[None, None], y_start_stop=y_start_stop,
xy_window=(96, 96), xy_overlap=(0.5, 0.5))
hot_windows = search_windows(image, windows, svc, X_scaler, color_space=color_space,
spatial_size=spatial_size, hist_bins=hist_bins,
orient=orient, pix_per_cell=pix_per_cell,
cell_per_block=cell_per_block,
hog_channel=hog_channel, spatial_feat=spatial_feat,
hist_feat=hist_feat, hog_feat=hog_feat)
window_img = draw_boxes(draw_image, hot_windows, color=(0, 0, 255), thick=6)
#window_img=cv2.cvtColor(window_img, cv2.COLOR_BGR2RGB)
window_list.append(window_img)
#plt.imshow(window_img)
plt.figure(figsize=(20, 20))
for i in range(len(window_list)):
plt.subplot(3,3,i+1)
plt.axis('off')
#axs[i].set_title('car', fontsize=10)
plt.imshow(window_list[i])
plt.savefig('./output_images/sliding_window_example.jpg')
More efficient method for sliding window¶
In [20]:
def convert_color(img, conv='RGB2YCrCb'):
if conv == 'RGB2YCrCb':
return cv2.cvtColor(img, cv2.COLOR_RGB2YCrCb)
if conv == 'BGR2YCrCb':
return cv2.cvtColor(img, cv2.COLOR_BGR2YCrCb)
if conv == 'RGB2LUV':
return cv2.cvtColor(img, cv2.COLOR_RGB2LUV)
# Define a single function that can extract features using hog sub-sampling and make predictions
# Both extract features and make predictions.
# It only has to extract hog features once and then can be sub-sampled to get all of its overlaying windows.
def find_cars(img, ystart, ystop, scale, svc, X_scaler, orient, hog_channel, pix_per_cell, cell_per_block, spatial_size, hist_bins):
draw_img = np.copy(img)
# Normalize
img = img.astype(np.float32)/255
img_tosearch = img[ystart:ystop,:,:]
ctrans_tosearch = convert_color(img_tosearch, conv='RGB2YCrCb')
if scale != 1:
imshape = ctrans_tosearch.shape
ctrans_tosearch = cv2.resize(ctrans_tosearch, (np.int(imshape[1]/scale), np.int(imshape[0]/scale)))
ch1 = ctrans_tosearch[:,:,0]
ch2 = ctrans_tosearch[:,:,1]
ch3 = ctrans_tosearch[:,:,2]
# Define blocks and steps as above
nxblocks = (ch1.shape[1] // pix_per_cell) - cell_per_block + 1
nyblocks = (ch1.shape[0] // pix_per_cell) - cell_per_block + 1
nfeat_per_block = orient*cell_per_block**2
# 64 was the orginal sampling rate, with 8 cells and 8 pix per cell
window = 64
nblocks_per_window = (window // pix_per_cell) - cell_per_block + 1
cells_per_step = 2 # Instead of overlap, define how many cells to step
nxsteps = (nxblocks - nblocks_per_window) // cells_per_step
nysteps = (nyblocks - nblocks_per_window) // cells_per_step
# Compute individual channel HOG features for the entire image
if hog_channel == "ALL":
hog1 = get_hog_features(ch1, orient, pix_per_cell, cell_per_block, feature_vec=False)
hog2 = get_hog_features(ch2, orient, pix_per_cell, cell_per_block, feature_vec=False)
hog3 = get_hog_features(ch3, orient, pix_per_cell, cell_per_block, feature_vec=False)
else:
hogx = get_hog_features(ctrans_tosearch[:,:,hog_channel], orient, pix_per_cell, cell_per_block, feature_vec=False)
# bbox
box = []
for xb in range(nxsteps):
for yb in range(nysteps):
ypos = yb*cells_per_step
xpos = xb*cells_per_step
# Extract HOG for this patch
if hog_channel == "ALL":
hog_feat1 = hog1[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel()
hog_feat2 = hog2[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel()
hog_feat3 = hog3[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel()
hog_features = np.hstack((hog_feat1, hog_feat2, hog_feat3))
else:
hog_features = hogx[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel()
xleft = xpos*pix_per_cell
ytop = ypos*pix_per_cell
# Extract the image patch
subimg = cv2.resize(ctrans_tosearch[ytop:ytop+window, xleft:xleft+window], (64,64))
# Get color features
spatial_features = bin_spatial(subimg, size=spatial_size)
hist_features = color_hist(subimg, nbins=hist_bins)
# Scale features and make a prediction
test_features = X_scaler.transform(np.hstack((spatial_features, hist_features, hog_features)).reshape(1, -1))
#test_features = X_scaler.transform(np.hstack((shape_feat, hist_feat)).reshape(1, -1))
test_prediction = svc.predict(test_features)
if test_prediction == 1:
xbox_left = np.int(xleft*scale)
ytop_draw = np.int(ytop*scale)
win_draw = np.int(window*scale)
cv2.rectangle(draw_img,(xbox_left, ytop_draw+ystart),(xbox_left+win_draw,ytop_draw+win_draw+ystart),(0,0,255),6)
box.append(((xbox_left, ytop_draw+ystart),(xbox_left+win_draw,ytop_draw+win_draw+ystart)))
return draw_img, box
test efficient sliding window¶
In [21]:
#y_start_stop = 400, 656
ystart, ystop = 400, 656
# It's possible to run this same function multiple times for different scale values
# to generate multiple-scaled search windows.
scale = 1.5
images = glob.glob('test_images/*.jpg')
test_list = []
box_list = []
for file_image in images:
image=cv2.imread(file_image)
img=cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
#img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
out_img, box = find_cars(img, ystart, ystop, scale, svc, X_scaler, orient, hog_channel, pix_per_cell, cell_per_block, spatial_size, hist_bins)
test_list.append(out_img)
box_list.append(box)
#plt.imshow(out_img)
plt.figure(figsize=(20, 20))
for i in range(len(test_list)):
plt.subplot(3,3,i+1)
plt.title('test_image'+str(i+1))
plt.axis('off')
#axs[i].set_title('car', fontsize=10)
plt.imshow(test_list[i])
plt.savefig('./output_images/advanced_sliding_window_example.jpg')
In [22]:
# Take an single image as an example
#scale = 1.5
single_img = mpimg.imread('test_images/test5.jpg')
single_out_img, single_boxs = find_cars(single_img, ystart, ystop, scale, svc, X_scaler, orient, hog_channel, pix_per_cell, cell_per_block, spatial_size, hist_bins)
plt.figure(figsize=(10,10))
plt.imshow(single_out_img)
plt.savefig('./output_images/sliding_window_example_test5.jpg')
Multiple Detections & False Positives¶
heat map
In [24]:
def add_heat(heatmap, bbox_list):
# Iterate through list of bboxes
for box in bbox_list:
# Add += 1 for all pixels inside each bbox
# Assuming each "box" takes the form ((x1, y1), (x2, y2))
heatmap[box[0][1]:box[1][1], box[0][0]:box[1][0]] += 1
# Return updated heatmap
return heatmap# Iterate through list of bboxes
def apply_threshold(heatmap, threshold):
# Zero out pixels below the threshold
heatmap[heatmap <= threshold] = 0
# Return thresholded map
return heatmap
def draw_labeled_bboxes(img, labels):
# Iterate through all detected cars
for car_number in range(1, labels[1]+1):
# Find pixels with each car_number label value
nonzero = (labels[0] == car_number).nonzero()
# Identify x and y values of those pixels
nonzeroy = np.array(nonzero[0])
nonzerox = np.array(nonzero[1])
# Define a bounding box based on min/max x and y
bbox = ((np.min(nonzerox), np.min(nonzeroy)), (np.max(nonzerox), np.max(nonzeroy)))
# Draw the box on the image
cv2.rectangle(img, bbox[0], bbox[1], (0,0,255), 6)
# Return the image
return img
In [26]:
# Read in image similar to one shown above
image = mpimg.imread('test_images/test5.jpg')
heat = np.zeros_like(image[:,:,0]).astype(np.float)
# Add heat to each box in box list
heat = add_heat(heat, single_boxs)
# # Apply threshold to help remove false positives
# heat = apply_threshold(heat,3)
# Visualize the heatmap when displaying
heatmap = np.clip(heat, 0, 255)
# Find final boxes from heatmap using label function
labels = label(heatmap)
draw_img = draw_labeled_bboxes(np.copy(image), labels)
print('cars found: ', labels[1])
fig = plt.figure(figsize=(20,20))
plt.subplot(121)
plt.imshow(draw_img)
plt.title('Car Positions', fontsize=20)
plt.subplot(122)
plt.imshow(heatmap, cmap='hot')
plt.title('Heat Map', fontsize=20)
fig.tight_layout()
plt.savefig('./output_images/heat_map_no_thresh_test5.jpg')
In [27]:
# Read in image similar to one shown above
thresh = 1
print('apply threshold: ', thresh)
image = mpimg.imread('test_images/test5.jpg')
heat = np.zeros_like(image[:,:,0]).astype(np.float)
# Add heat to each box in box list
heat = add_heat(heat, single_boxs)
# Apply threshold to help remove false positives
heat = apply_threshold(heat,thresh)
# Visualize the heatmap when displaying
heatmap = np.clip(heat, 0, 255)
# Find final boxes from heatmap using label function
labels = label(heatmap)
draw_img = draw_labeled_bboxes(np.copy(image), labels)
print('cars found: ', labels[1])
fig = plt.figure(figsize=(20,20))
plt.subplot(121)
plt.imshow(draw_img)
plt.title('Car Positions', fontsize=20)
plt.subplot(122)
plt.imshow(heatmap, cmap='hot')
plt.title('Heat Map', fontsize=20)
fig.tight_layout()
plt.savefig('./output_images/heat_map_with_thresh_test5.jpg')
In [28]:
def process_image(image):
'''process each single image'''
ystart, ystop = 400, 656
thresh = 1
scale = 1.5
# Converts color
img=cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
draw_image = np.copy(img)
heat = np.zeros_like(image[:,:,0]).astype(np.float)
# Finding car positions
# Note normalization function has been writen in the find_cars
_, single_boxs = find_cars(img, ystart, ystop, scale, svc, X_scaler, orient, hog_channel, pix_per_cell, cell_per_block, spatial_size, hist_bins)
# Add heat to each box in box list
heat = add_heat(heat, single_boxs)
# Apply threshold to help remove false positives
heat = apply_threshold(heat, thresh)
# Find final boxes from heatmap using label function
labels = label(heat)
draw_img = draw_labeled_bboxes(draw_image, labels)
return draw_img
In [29]:
#y_start_stop = 400, 656
ystart, ystop = 400, 656
# It's possible to run this same function multiple times for different scale values
# to generate multiple-scaled search windows.
scale = 1.5
images = glob.glob('test_images/*.jpg')
test_list = []
box_list = []
for file_image in images:
image=cv2.imread(file_image)
#img=cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
#image = image.astype(np.float32)/255
draw_img = process_image(image)
test_list.append(draw_img)
box_list.append(box)
#plt.imshow(out_img)
plt.figure(figsize=(20, 20))
for i in range(len(test_list)):
plt.subplot(3,3,i+1)
plt.title('test_image'+str(i+1))
plt.axis('off')
#axs[i].set_title('car', fontsize=10)
plt.imshow(test_list[i])
plt.savefig('./output_images/heat_map_with_thresh_examples.jpg')
define class¶
In [30]:
# Define a class to receive the characteristics of each vehicle detection
class Detection():
def __init__(self, n):
# How many frames to use to smooth
self.n = n
# Line smoothing
self.windows = []
def add_to_smooth_detect(self, window):
'''Smoothing over the last n frames'''
self.windows.append(window)
# Keep size of windows equal to n
if len(self.windows)>self.n:
self.windows.pop(0)
#self.ave_A = np.mean(self.A)
return self.windows
def get_results_of_smooth_detect(self):
return self.windows
process each frame with smoothing¶
In [34]:
def process_each_frame(image):
'''process each frame of the video'''
# ystart, ystop = 400, 656
# thresh = 1
# scale = 1.5
# Converts color
#img=cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
draw_image = np.copy(image)
heat = np.zeros_like(image[:,:,0]).astype(np.float)
# Finding car positions
# Note normalization function has been writen in the find_cars
_, single_boxs = find_cars(image, ystart, ystop, scale, svc, X_scaler, orient, hog_channel, pix_per_cell, cell_per_block, spatial_size, hist_bins)
# Smooth frames
if len(single_boxs)>0:
detection.add_to_smooth_detect(single_boxs)
# Return previous windows
previous_windows = detection.get_results_of_smooth_detect()
# Add heat to each box in box list
for window in previous_windows:
heat = add_heat(heat, window)
#print(len(previous_windows))
# Apply threshold to help remove false positives
heat = apply_threshold(heat, 1+len(previous_windows)//2)
# Find final boxes from heatmap using label function
labels = label(heat)
draw_img = draw_labeled_bboxes(draw_image, labels)
return draw_img
In [60]:
ystart, ystop = 400, 656
#thresh = 1
scale = 1.5
frames = 10
detection = Detection(n=frames)
img = mpimg.imread('test_images/test6.jpg')
draw_img = process_each_frame(img)
plt.imshow(draw_img)
plt.savefig('./output_images/final_example_test5.jpg')
output video¶
In [38]:
# Import everything needed to edit/save/watch video clips
from moviepy.editor import VideoFileClip
from IPython.display import HTML
ystart, ystop = 400, 656
#thresh = 1
scale = 1.5
frames = 10
detection = Detection(n=frames)
test_out_file = 'project_video_result.mp4'
clip_test = VideoFileClip('project_video.mp4')
clip_test_out = clip_test.fl_image(process_each_frame)
%time clip_test_out.write_videofile(test_out_file, audio=False)
Combine with detecing lines¶
In [66]:
import numpy as np
import cv2
import pickle
import glob
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
from moviepy.editor import VideoFileClip
from IPython.display import HTML
import pickle
from Line_class import Line
from combined_thresh import combined_thresh
from Perspective_transform import perspective_transorm
from polynomial_fit import line_fit, advanced_fit, line_fit_visualize, advanced_fit_visualize
% matplotlib inline
def calculate_curvature(leftx, rightx, lefty, righty):
'''Calculate the radius of curvature in meters'''
# Define y-value where we want radius of curvature
# I'll choose the maximum y-value, corresponding to the bottom of the image
#y_eval = np.max(ploty)
y_eval = 719
# Define conversions in x and y from pixels space to meters
ym_per_pix = 30/720 # meters per pixel in y dimension
xm_per_pix = 3.7/700 # meters per pixel in x dimension
# Fit new polynomials to x,y in world space
left_fit_cr = np.polyfit(lefty*ym_per_pix, leftx*xm_per_pix, 2)
right_fit_cr = np.polyfit(righty*ym_per_pix, rightx*xm_per_pix, 2)
# Calculate the new radii of curvature
left_curverad = ((1 + (2*left_fit_cr[0]*y_eval*ym_per_pix + left_fit_cr[1])**2)**1.5) / np.absolute(2*left_fit_cr[0])
right_curverad = ((1 + (2*right_fit_cr[0]*y_eval*ym_per_pix + right_fit_cr[1])**2)**1.5) / np.absolute(2*right_fit_cr[0])
# Now our radius of curvature is in meters
return left_curverad, right_curverad
def calculate_offset(undist, left_fit, right_fit):
'''Calculate the offset of the lane center from the center of the image'''
xm_per_pix = 3.7/700 # meters per pixel in x dimension
ploty = undist.shape[0] # height
left_fitx = left_fit[0]*ploty**2 + left_fit[1]*ploty + left_fit[2]
right_fitx = right_fit[0]*ploty**2 + right_fit[1]*ploty + right_fit[2]
offset = (left_fitx+right_fitx)/2 - undist.shape[1]/2 # width
offset = xm_per_pix*offset
return offset
def final_drawing(undist, left_fit, right_fit, left_curverad, right_curverad, Minv, vehicle_offset):
'''Project the measurement back down onto the original undistorted image of the road'''
# Generate x and y values for plotting
ploty = np.linspace(0, undist.shape[0]-1, undist.shape[0] )
left_fitx = left_fit[0]*ploty**2 + left_fit[1]*ploty + left_fit[2]
right_fitx = right_fit[0]*ploty**2 + right_fit[1]*ploty + right_fit[2]
# Create an image to draw the lines on
warped = np.zeros((720,1280))
warp_zero = np.zeros_like(warped).astype(np.uint8)
color_warp = np.dstack((warp_zero, warp_zero, warp_zero))
# Recast the x and y points into usable format for cv2.fillPoly()
pts_left = np.array([np.transpose(np.vstack([left_fitx, ploty]))])
pts_right = np.array([np.flipud(np.transpose(np.vstack([right_fitx, ploty])))])
pts = np.hstack((pts_left, pts_right))
# Draw the lane onto the warped blank image
cv2.fillPoly(color_warp, np.int_([pts]), (0,255, 0))
# Warp the blank back to original image space using inverse perspective matrix (Minv)
newwarp = cv2.warpPerspective(color_warp, Minv, (undist.shape[1], undist.shape[0]))
# Combine the result with the original image
result = cv2.addWeighted(undist, 1, newwarp, 0.3, 0)
# Anotate curvature values
ave_curvature = (left_curverad + right_curverad)/2
ave_text = 'Radius of average curvature: %.2f m'%ave_curvature
cv2.putText(result, ave_text,(50,50), 0, 1, (0,0,0), 2, cv2.LINE_AA)
# Anotate vehicle offset from the lane center
if(vehicle_offset>0):
offset_text = 'Vehicle right offset from lane center: {:.2f} m'.format(vehicle_offset)
else:
offset_text = 'Vehicle left offset from the lane center: {:.2f} m'.format(-vehicle_offset)
cv2.putText(result, offset_text,(50,80), 0, 1, (0,0,0), 2, cv2.LINE_AA)
#plt.imshow(result)
return result
def process_video_image(img):
'''Process each image in the video and return with annotated image'''
global detected, mtx, dist, left_lanes, right_lanes
# Undistorting image
undist = cv2.undistort(img, mtx, dist, None, mtx)
# Combined thresh
combined = combined_thresh(undist)
# Warped image
binary_warped, _, Minv = perspective_transorm(combined)
# Line fit
if not detected:
curv_pickle = line_fit(binary_warped)
if curv_pickle is not None:
left_fit = curv_pickle["left_fit"]
right_fit = curv_pickle["right_fit"]
leftx = curv_pickle["leftx"]
lefty = curv_pickle["lefty"]
rightx = curv_pickle["rightx"]
righty = curv_pickle["righty"]
# Update
# Add into the set to smoothing average fit
left_fit = left_lanes.add_to_smooth_fit(left_fit)
right_fit = right_lanes.add_to_smooth_fit(right_fit)
# Calculate curvature
# Assume first frame can be detected both lanes and curvatures
left_curvature, right_curvature = calculate_curvature(leftx, rightx, lefty, righty)
# # Add into the set to smoothing average curvature
# # Only use for a frame without detected lanes
# special_left_curv = left_lanes.ave_curvature(left_curvature)
# special_right_curv = right_lanes.ave_curvature(right_curvature)
detected = True
else:
# Calculate based on previous frames
left_fit = left_lanes.get_results_of_smooth_fit()
right_fit = right_lanes.get_results_of_smooth_fit()
special_left_curv = left_lanes.get_results_of_ave_curvature()
special_right_curv = right_lanes.get_results_of_ave_curvature()
left_curvature, right_curvature = special_left_curv, special_right_curv
detected = False
else:
# Smooth fit
left_fit = left_lanes.get_results_of_smooth_fit()
right_fit = right_lanes.get_results_of_smooth_fit()
# Add into the set to smoothing average curvature
# Only use for a frame without detected lanes
special_left_curv = left_lanes.get_results_of_ave_curvature()
special_right_curv = right_lanes.get_results_of_ave_curvature()
# Skip the sliding windows step once you know where the lines are
# Search in a margin around the previous line position
curv_pickle = advanced_fit(binary_warped, left_fit, right_fit)
if curv_pickle is not None:
# Detected lines
left_fit = curv_pickle["left_fit"]
right_fit = curv_pickle["right_fit"]
leftx = curv_pickle["leftx"]
lefty = curv_pickle["lefty"]
rightx = curv_pickle["rightx"]
righty = curv_pickle["righty"]
# Note: We only make update when detect lanes in the current frame
# Add into the set to smoothing average fit
left_fit = left_lanes.add_to_smooth_fit(left_fit)
right_fit = right_lanes.add_to_smooth_fit(right_fit)
# Calculate curvature
left_curvature, right_curvature = calculate_curvature(leftx, rightx, lefty, righty)
else:
# Calculate curvature based on previous frames
left_curvature, right_curvature = special_left_curv, special_right_curv
detected = False
# Calculate vehicle offset from the lane center in the image
vehicle_offset = calculate_offset(undist, left_fit, right_fit)
if vehicle_offset > 0.25:
detected = False
# Project the measurement back down onto the original undistorted image of the road
result = final_drawing(undist, left_fit, right_fit, left_curvature, right_curvature, Minv, vehicle_offset)
return result
def final_process_image(img):
result = process_video_image(img)
return process_each_frame(result)
if __name__ == '__main__':
# Read in the saved camera matrix and distortion coefficients
dist_pickle = pickle.load(open("./camera_cal/camera_dist_pickle.p", "rb"))
mtx = dist_pickle["mtx"]
dist = dist_pickle["dist"]
# Define number of frames for smoothing
num_frames = 5
left_lanes = Line(n=num_frames)
right_lanes = Line(n=num_frames)
detected = False
img = mpimg.imread('./test_images/test6.jpg')
#result = process_video_image(img)
result = final_process_image(img)
#plt.imshow(result)
plt.savefig('./output_images/final_combined_example_test5.jpg')
# mpimg.imsave('./output_img/test2_final_projected.jpg', result)
# # Visualize images
# f, (ax1, ax2) = plt.subplots(1, 2, figsize=(24, 9))
# f.tight_layout()
# ax1.imshow(img)
# ax1.set_title('Original Image', fontsize=50)
# ax2.imshow(result)
# ax2.set_title('Polynomial Fit Image', fontsize=50)
# plt.subplots_adjust(left=0., right=1, top=0.9, bottom=0.)
# plt.savefig('./output_img/test2_final_projected_example.jpg')
# # Import everything needed to edit/save/watch video clips
# from moviepy.editor import VideoFileClip
# from IPython.display import HTML
# Define number of frames for smoothing
num_frames = 8
left_lanes = Line(n=num_frames)
right_lanes = Line(n=num_frames)
detected = False
ystart, ystop = 400, 656
#thresh = 1
scale = 1.5
frames = 10
detection = Detection(n=frames)
white_output = 'test_demo.mp4'
clip1 = VideoFileClip("project_video.mp4")
white_clip = clip1.fl_image(final_process_image) #NOTE: this function expects color images!!
#%time white_clip.write_videofile(white_output, audio=False)
white_clip.write_videofile(white_output, audio=False)
In [70]:
ystart, ystop = 400, 656
#thresh = 1
scale = 1.3
frames = 10
detection = Detection(n=frames)
white_output = 'test_demo——s.mp4'
clip1 = VideoFileClip("test_video.mp4")
white_clip = clip1.fl_image(final_process_image) #NOTE: this function expects color images!!
#%time white_clip.write_videofile(white_output, audio=False)
white_clip.write_videofile(white_output, audio=False)
In [ ]: